Alternator transformer/rectifier remote mounted (test 1.0)

[mobile_bob]

i decided to chronicle this project somewhere apart from the white paper section, as it is a work in progress
and not a finished project by a stretch

overview of this project:

the goal of this project is to use an otherwise stock oem alternator, and to take three phase power off the stator at a higher voltage
than would be normal, transfer the generated AC over a 10/3 stranded cable some distance. this 3 phase higher voltage power would then be
stepped down to approx 12vac line to line  using a transformer pack made up of 3 step down transformer connected appropriately to attain the
requisite 12vac, which will then be fed to the remote rectifier which will convert the 3 phase power to DC which will be used to charge a 12volt battery bank.

what i would like to determine is what is the relative efficiency of this system compared to a stock alternator charging a 12volt battery at normal distances and also how it compares in cost and efficiency with a stock 12volt alternator that must charge a battery mounted some 30 ft or more away.

there are lots of unknowns going in on this project, and i am going on what i have learned from research on the lundell/clawpole automotive alternator and the use of transformers that were made for 60 hz use being repurposed for use at 500-600hz.

not wanting to spend money on a project of unknown odds of success, i have chosen to use surplus parts to do the testing, and use them to
prove the concept, at the expense of having a useful operational system perhaps. if the theory of operation is good, and if what i have been told by transformer manufactures EE is correct, then the preliminary testing should prove the concept and then taking what is learned perhaps i will choose to either alter certain components or buy the ones needed to get the voltage within the needed operating range.

what i have to work with

a leece neville 4400 series 270amp alternator, @12 volts. it has the stator leads accessible from the rear of the case, and i have a remote
regulator plate on the alternator so that i have access directly to the brushes.

the transformers i am using are from APC rack mounted 1kva ups systems, i will be using three of them that have 120 volt primaries
and ~24volt secondaries

the rectifier bridges are a pair of negative and positive sets from a salvaged 175amp leece neville


the hookup

i connected a 10ft piece of 10/3 stranded cable to the stator posts at the back of the alternator, then connected the three transformer primaries
in a delta configuration, and attached the other end of the 10/3 to the transformer bank

the secondaries of the transformers i chose to connect in Y configuration, so as to get max voltage from the step down of the trans set
figuring if it were too high i could reconnect as delta to decrease voltage and increase amps later.

the Y connected secondaries i connected to the rectifier bridge, and left the rectifier open circuit for preliminary testing

test one:

i brought the plant up to rated engine speed of 1800rpm, the indicated voltage was ~40vac phase to phase coming off the stator

i would have liked to have higher voltage but settled with this for preliminary testing as i had to full field the alternator to get this much
AC voltage off the stator, the stator is internally delta connected, so later should i choose to change the stator to Y connection i could get approx
40 x 1.73 = 69.2 volts (which is probably more in line with what parts i have to work with actually need)

i checked the line current to the transformer and found it to not even read on my clamp on meter, i would assume there to be a bit of current flowing but because of the current being ~500hz the meter might not be able to accurately measure that level of power. i did not that the engine
does not change tone with power to the transformer primaries nor does the transformer cores even warm up,, so the power level used to magnetize the cores and cover core losses might be pretty low.

i measured the voltage off the rectifier bridge and it came in at ~11 volts, not high enough to charge a 12volt battery, but this is a test and i am not that lucky to have parts that work perfectly for every project.

i then placed my carbon pile on the DC output and cranked up a 50amp load, the voltage dropped to approx 4 volts, and the stator voltage dropped to about 37volts, here having a 69volt input would be handy as there would be ample overhead to cover the drop under load and not have to full field the alternator,

i checked the stator current to the primaries and found it to be ~9.5-10.0 amps with a 50amp load, here again i expect the accuracy might be off a bit at the higher hz,

by applying the 50amp load and taking it away againl, off and on, i could not tell any change in tone of the engine, it was as if it didn't see a load at all, however 50amps and 4volts is only a 200watts, even with efficiency hits of conversion it is unlikely the load would be much more than 300 watts as seen by the engine, and i am not sure i could hear a tone change with 300 watts or not.

what i have learned so far

1. i can generate higher stator voltages, (actually i have known this for sometime) and i need to reconnect a stator as Y configuration
so that i might see around 70vac phase to phase at 1800rpm engine speed (approx 5000rpm alternator speed), or
i need to change my drive ratio and increase the alternator speed or maybe do both

2. the core losses of the transformers seem quite acceptable, i was expecting to see some heating of the cores being run at 8 to 10 times
their normal 60hz operation,

3. i need to change the transformer turns ratio, in lieu of changing #1 above, that would be very easy to do with these transformers as the secondaries are wound over the primaries and are accessible so i can remove a few turns, there are 24 turns now and reducing the turn count to
16 likely would put me close to what i need, and possibly increasing the alternator rpm might get me into the sweet spot of about 50% excitation current at moderate charge current.

CORRECTION:   reducing the secondary turn count will further reduce my needed voltage, so that is not a viable option, so disregard #3 above
                      it would appear i will be needing to confine the alterations to changing the alternator stator from delta to Y configuration, or
                      increasing alternator speed, and maybe both

this first test looks encouraging, and promising
it would appear that i have achieved proof of concept, and there does not appear to be any major issues standing in the way of success

i looks like the next step might be to go ahead and choose the easiest to modify change, and then run another battery of tests to determine the outcome so that i can see what other tweaks will be necessary to dial this system in.

once i have the system to where it will provide charging at ~100amps continuous at 14.4 volts, thermally stable at both the alternator and the
remote transformers and rectifier, then the final battery of tests will be comparative to determine the relative efficiency of this system compared to a stock oem alternator with batteries closely located, and also compared against a stock alternator charging a set of batteries at some longer distance such as 30 ft.

in theory the system ought to be about even in efficiency for close charging and perhaps higher in efficiency for distance charging, and it is possible it will be more efficient than either method.

well there you guys have it, the first testing i have been able to do in over 6 months, and it feels damn good to be able to do it again.



bob g

[mobile_bob]

after a bit more thought, i think i will do the following

i have one more of these transformers, i think i will look into carefully taking it apart
and removing turns from the primary so that i might get the output voltage from the secondary that i need

the primary is designed for 120vac anyway, and i am down to around 40vac so i could probably take off enough turns
to get my secondary voltage up to where i need it and likely improve the transformers efficiency anyway.

i think i can remove turns by uncoiling them from the center easy enough, rewrap the set and dip the finished coil
then reassemble and drive in wood wedges to secure the coil to the core.

that might be a reasonable approach, gotta think about this a bit more.

i suppose another approach might be removing the secondaries and rewinding them with more turns of 3 or 4 in hand gage wire
to get what i need as well, i might be able to do that without disassembling the core.

yes more options

bob g

[mobile_bob]

well i took the option of reducing turns on the primaries, and decided to use my spare transformer as a guinea pig

lesson learned, there are about a million laminations in one of these transformers, E/I laminations about .008" thick which is about
half the thickness i was expecting.

took all of an hour to seperate the laminations, and remove the primary winding, it is very thick rectangular section and came off like a big
soft spring, so when i am done with the primary alteration it should spool back on without a lot of problems.

i think i will finish the mods to this transformer, and see how it works and how much more time i will have invested in it, i am thinking probably 2.5 hours per transformer will be about right,, and as far as i am concerned it doesn't make economic sense to do so.

somehow this is much more difficult that i remember it being when i was a kid?

of course everything cost less, ran faster, was easier and bigger all those years ago, from memory



bob g

[BruceM]

Great progress, Bob!  It could be worse- if the transformers were toroids, imagine the fun unwinding and rewinding each turn.  I can't redo a primary on a toroid anymore.

I hope your transformer primary mod. goes smoothly and you get your voltage right on.  An extra tap isn't cheating, if you have room for the extra loop leadout.

I hope you won't pay later for your project fun today!

[mobile_bob]

Bruce:

i can't even imagine rewinding a toroidal transformer  

as it is, it took me about 3.5 hours to comeplete the rewind of the primary of this unit
and i am now faced with doing 2 more to get a set for the 3 phase project.

good lord 10.5 hours for a set of three transformers is either a labor of love or insanity
i am thinking it is the latter.

it went surprisingly smoothly after i decided on the amount of turns i should remove from the primary
and yes i considered using taps, and probably should have!

if nothing else taps would have made the unit useful as a battery charger for single phase source had I decided not
to rewind two more.

i guess i can justify the time requirement as being an educational event, yes that is what i will tell myself!

later this morning i need to dig out my variac and apply 30-40 volts to the new primary and see how many volts i can get off the secondaries now
i am shooting for 10.6 vac with approx 35 vac input.

hopefully i am close, i think i should be or rather

i will always think it should have been right!!



at this point i think it might have been much easier to reconnect the stator of the alternator as Y from delta, i have an extra 175amp stator
and the way they made it, it can easily be connect either delta or Y, because they attached ring terminals to each end of each coil and pair them up to the stator posts for connection to the rectifier.

i think before i invest more time in rewinding the other two transformers i may well reconnect the stator and do another test run to see how things play out, i suspect that would be a much more expedient means to an acceptable end, given i am not looking for a very high power use, but rather an outline on how to accomplish this goal , as related by our friend Lloyd.

his need for the transfer of large amount of power over 30ft for 12volt battery charging is something that i would find an easier work around, like
moving the batteries much closer to the charging source, however i realize he cannot do this and i suspect his application might not be the only one where such distances are involved.

all said and done, with some good testing to document the results, this work will end up as an addition to the white paper already published here, and in the end the hope would be, we all should have a very good idea of what can, cannot, and maybe even what should not be done with automotive alternators.

it feels really good to have felt good enough to actually get something done out in the shop sunday, it was the first time i felt able since about june of last year.

my how time flies when you feel like crap!

thanks Bruce for your comments, it is always appreciated

bob g

[mobile_bob]

a working sketch of the system

bob g

[BruceM]

I agree, Bob, it sounds like you'd be better off converting the alternator end for testing purposes.  Over ten hours of rewinding  is a bit much! 

[mobile_bob]

test #2

i opted to replace the stator in my test mule (leece neville 270amp) for a 175amp stator
reconfigured for Y from the stock config of delta

the open circuit phase to phase voltage came up to 99vac at full field and the DC side settled in at ~25vdc

this gives me plenty of headroom to work with, by application of load and regulation of the field down from full field.

the quick and dirty load test returned the following numbers

DC output at full field and 100amps @ 13.5vdc

AC voltage at full load 90vac, AC current phase per phase 18amps

now we are starting to get dialed into the ballpark of what is needed.

transformer presented no odd sounds, no smoke and no heating, and the 10 gage 3 conductor SO cable looks to be adequate to
transfer this level of power.

i think i could pull significantly more current from this setup, but my old carbon pile needs serviced and it is not really the right tool for continous
load testing, in later testing i will insert a fixed resistive load that will handle a continuous duty requirement.

the next step is cleaning up the rather crude breadboarding to the point that i can apply a fixed load and do some comparative efficiency
tests

so far i am getting more and more encouraged, it is pretty cool seeing 100amps being transferred on the equivalent of an extension cord
and have 13.5 vdc which is adequate for this test to actually charge a battery, the voltage climbs well as the load is decreased. i plan on
producing a graph showing amps vs voltage soon as i get things ironed out a bit more. i also did not notice any change in engine sound indicating
and increased load, which tells me the efficiency ought to be pretty good, 13.5 x 100amps is a 1350watt load plus transformer losses, rectifier losses and alternator losses, i would have expected to hear the engine change tone if the overall efficiency was typical of an automotive alternator
used as designed. if the system was 50% efficient as automotive alternators often are, then to have 1350 watts at the user end i would have to produce 2700 watts of power from the engine,, that would cause a decided change of tone with the engine.

going out on a limb here, i think the system might be at least as efficient as an oem alternator, and not take a hit for the transformers.

i know that is not very scientific, but i see it as an indicator that the path i am on is getting a bit smoother.

more to come

this is getting to be very interesting indeed

on a side note,  i think these apc ups transformers are very high quality, and my bet is one could build a nice battery charger with one or two
for an st head, or build a nice 120 or 240volt welder capable of doing some decent work probably adequate for most diy'ers projects.

bob g

[BruceM]

Wow, Bob, that was fast work on the alternator delta wiring mod and retest.  Looking forward to seeing how your efficiency data turns out. 

[mobile_bob]

this particular series and type of alternator makes changing from delta to Y a breeze

both end of each of the three phase coil have their own eyelet soldered on, so you can connect
them as you wish if you are careful to insulate, tuck and secure the neutral connection of the Y

all told with R/R time probably less than a half hour, and i was up and ready to run again.

truth be told i needed to make the modification anyway for some other testing i wanted to do, and i got
the info i needed from this #2 test that will aid me in determining what the outcome is likely to be for that test.

funny how more pieces of the puzzle (data points) make predicting outcomes much more accurate an endeavor.

bob g

[Geno]

Bob, are you using your Balmar to vary the field current? If not, what?  I'm up to that stage in my battery charger project. I did some testing last night with mixed results. I'll post them on my thread when I sort things out a little more.

Thanks, Geno

[mobile_bob]

yes i am currently using a controller that is similar in function but made or marketed by Hehr

it does the same thing, but has pots to adjust rather than programming,

in the end i will probably use another balmar, but don't like risking one in development stages of a project.

bob g

[mike90045]

and the 10 gage 3 conductor SO cable looks to be adequate to transfer this level of power.

Most data indicates 10ga wire is safe at 30A,  if you are pushing 100A, (3x limit) I think you WILL get a suprise, unless it's on ceramic post & tube insulators. 

[BruceM]

It's three phase power, Mike, so each wire is carrying 1/3 the current- and that's the current at higher voltage, much less than 100amps.  That and the improved efficiency of the alternator at higher volts is the beauty of Bob's design.

[mobile_bob]

the 100amps comes after the three phase transfomer step down and rectifier bridge

basically this thing is a three phase battery charger being plugged into a 3 phase 500hz wall socket (the alternator)

just as a battery charger from sears might pump out 100amps to charge a battery, it does not draw 100amps from the wall socket
but a small fraction of that amount.

also after giving this more thought, and getting some guidance from my favorite EE forum, it would appear i am drawing ~2400watts
over the 10/3 line, but  a large percentage of that is reactive power which i need to address with capacitors because my powerfactor
is probably not at all good feeding transformers.

also 500hz is a difficult thing to measure amps with a clamp on meter, likely not very accurate, so

i will probably choose to push ahead to test #3 and just do comparative BSFC testing to see how the system compares with a stock charging
system, the reactive power for the most part will not be a factor and i can then deduce what the actual current is, and calculate the reactive power from that.

sort of a back door maneuver, this should give me a data set to compare with, so that in test #4 i can use Bill Rogers simple method of power factor correction, monitoring voltage for increase which i can do with a voltmeter no problem and not worry about measuring actual amperage.
(or reduction thereof)

got a ways to go yet,

bob g